TIOCGSID is used by tcgetsid() so it is not really compat :-)
This should reduce loading the compat module.

(christos)

Support using EFI runtime services for RTC when booting in devicetree mode.

(jmcneill)

Support early FB console attachment when booting with a devicetree
(non-ACPI mode). Inform the pciconf code about the framebuffer to
prevent pciconf from changing resources out from under us when framebuffer
memory is in VRAM.

(jmcneill)

Add pciconf_resource_reserve. This allows MD code to mark specific memory
and I/O regions as in use. When pciconf finds a device already configured
to use one of these regions, the device config is left as-is.

(jmcneill)

Ticket #1614

(martin)

Pull up following revision(s) (requested by nia in ticket #1614):

sys/compat/common/tty_43.c: revision 1.38

tty_43: Check a bitset from userspace is valid before shifting it

Passing a negative value to these legacy compat ioctls results in
left shift on a negative value which is undefined behaviour and results
in the tty (at least, possibly other things) locking up.
The argument to the ioctl should always be > 0. Return EINVAL otherwise.

While here, adjustments to code style to match current guidelines.
Found by UBSan.

(martin)

Ticket #1106

(martin)

Pull up following revision(s) (requested by nia in ticket #1106):

sys/compat/common/tty_43.c: revision 1.38

tty_43: Check a bitset from userspace is valid before shifting it

Passing a negative value to these legacy compat ioctls results in
left shift on a negative value which is undefined behaviour and results
in the tty (at least, possibly other things) locking up.
The argument to the ioctl should always be > 0. Return EINVAL otherwise.

While here, adjustments to code style to match current guidelines.
Found by UBSan.

(martin)

PR/55710: Kimmo Suominen: Delete manual pages so that they get copy from the
imported, already generated ones.

(christos)

Use .PARSEDIR

(christos)

Add rules for manual pages builds, remove including Makefile.inc

(christos)

include bsd.own.mk so that ../../Makefile.inc does not do it...

(christos)

tty: Negating INT_MIN will overflow int, bail out with EINVAL

Detected by UBSan

Reported-by: syzbot+92c0fca82b74a9798b78@syzkaller.appspotmail.com

(nia)

Wait for the asynchronous dns resolver child to be done (and close the
socket descriptors) before returning. Otherwise we might get bind errors.
Reported by kim@, fixed by mlelstv@.

(christos)

merge 20200606: resolve conflict by adding kb2 and rin entries to putty

(christos)

xeyes should be AlwaysOnTop - requested by a user

(nia)

Bump up the dark contrast a bit

(nia)

Make it easier to identify the current active window (red border highlight)

(nia)

Read the linux,pci-probe-only property from the /chosen node, not the PCI host controller node

(jmcneill)

Switch to the same hack in pkgsrc for avoiding bitmap fonts on the web

Just for consistency's sake, and making it easier to disable this if
someone wants to do that.

(nia)

hook up X11 "Spleen" fonts to the build

(nia)

Don't change a videomode that's been setup in the firmware.
Extend the register debugging output (and hopefully make it easier to read).

(jdc)

Report half-precision FP and SIMD support

(jmcneill)

Fix detection of NEON features. ID_AA64PFR0_EL1_ADV_SIMD_NONE means SIMD
is not available, and any other value means it is.

(jmcneill)

Fix detection of FP and SIMD features on Armv8.2+.

(jmcneill)

Add Makefile for X11 Spleen fonts. Not integrated yet.

(nia)

Fix MKDEBUG=yes build for hppa.

(rin)

reorder functions for better relevance

(nisimura)

Fix cycle counter-based time keeping on Alpha in MP environments by using
a simpler calibration algorithm for the CC timecounter.  Proposed in 2018
by Naruaki Etomi:

    https://mail-index.netbsd.org/tech-kern/2018/01/14/msg022940.html

This patch is largely based on the proposed change, but avoids changing
any other timecounter logic, and re-factors things a bit to keep them
as MI as possible.

(thorpej)

Use correct note types for register storage in 32-bit core files for
architecture on which 64- and 32-bit ABIs use different values for
PT_GET{,FP}REGS, i.e., aarch64{,eb}.

Now, 32-bit GDB works fine for core files generated by aarch64{,eb}
kernel.

Should be no functional changes for ports other than aarch64{,eb}.

(rin)

regen for getrandom(2)

(rin)

Teach about getrandom(2) to COMPAT_NETBSD32.

(rin)

Build GDB for aarch64eb; it works just fine as far as I can see.

(rin)

Adjust for aarch64eb; it needs elf32-aarch64.c.

(rin)

mknative-gdb.old for aarch64eb

(rin)

Correct target directory; src/external/gpl3/gdb.old not gdb.

(rin)

Nowadays, mknative-gdb puts files into external/gpl3/gdb/lib,
instead of external/gpl3/gdb/bin/gdb.

Also, we need to nbmake-MACHINE in external/gpl3/gdb in order to
build dependent libraries correctly.

(rin)

Don't do extra work finding the power of 2 for values we are not going to
use. Explain that cpu_hatch has not been called yet, so no cpu_probe either
so the cache info is 0 for AP's.

(christos)

Add italic correction to _Atomic-qualified.

(uwe)

new tzcode

(christos)

Merge tzcode2020b (except we keep tzsetwall(3) for now for compatibility,
and we were "slim" already)

Support for zic's long-obsolete '-y YEARISTYPE' option has been
removed and, with it, so has support for the TYPE field in Rule
lines, which is now reserved for compatibility with earlier zic.
These features were previously deprecated in release 2015f.
(Thanks to Tim Parenti.)

zic now defaults to '-b slim' instead of to '-b fat'.

zic's new '-l -' and '-p -' options uninstall any existing
localtime and posixrules files, respectively.

The undocumented and ineffective tzsetwall function has been
removed.

(christos)

When MD code requests a reserved area (at the start of the disk)
make sure to properly account for the "lost" space during all calculations
and especially when expanding the main partition to fill the disk.

Avoid using the global "pm" or a passed device descriptons when we are
passing disk limits explicitly anyway.

Simplify/fix a few rounding calculations.

(martin)

Add tests for x87 FPU registers

Reviewed by kamil

(mgorny)

Rename MM_REG macro to ST_MAN, and cover fpr/xstate with it

Rename the MM_REG macro to ST_MAN, to make it clearer that it gets
mantissa of ST registers which overlaps with MM registers but can be
also used to read ST registers (to be used in the next commit).  Extend
it to cover the difference between GETFPREGS and GETXSTATE,
and therefore avoid additional condition on i386.

Reviewed by kamil.

(mgorny)

Reword advice about when not to use membar_exit.

With help from skrll and pgoyette.

While here, change the example so it doesn't violate the advice just
given.  Still not a great example -- abusing UINT_MAX for a reference
count is kinda sketchy, but it'll serve for now.

(riastradh)

build ppc_video.c etc. on hppa

(macallan)

Oops, avoid shadow declaration (fixes pci_system_netbsd_count_buses)

(jmcneill)

tty_43: Check a bitset from userspace is valid before shifting it

Passing a negative value to these legacy compat ioctls results in
left shift on a negative value which is undefined behaviour and results
in the tty (at least, possibly other things) locking up.

The argument to the ioctl should always be > 0. Return EINVAL otherwise.

While here, adjustments to code style to match current guidelines.

Found by UBSan.

Reported-by: syzbot+39cd551a05298b222756@syzkaller.appspotmail.com

(nia)

Initial import of font-spleen-misc-1.8.2.

Discussed with nia@.

(fcambus)

tty: Avoid undefined behaviour (left shift of 1 by 31 places overflows int)

The valid sizes of the tty input and output queues (according to the man page)
are between 1024 and 65536 and input values are converted to a power of two.

The check on the validity of the range is done after the input values are
converted, however, which means that a hostile program can attempt to set
the queue size to a negative value, and cause integer overflow before
the range is validated.

Detected by UBSan

Reported-by: syzbot+521b73969fd233c49e58@syzkaller.appspotmail.com

(nia)

PR kern/55705:
don't attempt to sleep for negative time, we are late anyway - avoids DIAGNOSTIC panic

(kardel)

carp: Remove media, software should use link status.

carp literally has no media just like ppp, vlan, etc.

(roy)

make(1): add test for parsing the :? variable modifier

(rillig)

make(1): use consistent order in .OODATE in test archive.mk

(rillig)

switch powerpc, sh3, vax and m68k to GCC 9.  that's everyone.

thanks espcially to rin@ for testing and fixing issues.

(mrg)

set list for hppa

(macallan)

build some potentially useful drivers on hppa

(macallan)

build Xorg server on hppa

(macallan)

options.4: add a basic entry for the CNMAGIC option

(gutteridge)

Fix crash when more than 32 pci(4) devices are attached.

(jmcneill)

rk3399_pcie: do reset timing according to PCI Express Base Specification

Don't do link training or configuration space accesses within the time
allowed by the standard for the downstream card to come out of reset
after deasserting PERST#.

This fixes detection issues seen with a dual port wm(4) NIC,
an ASMedia SATA card and also Pericom bridges (but they need more work
to be useful).

(tnn)

add DM9000A register definitions

(nisimura)

make(1): remove test misc.mk from Makefile

(rillig)

make(1): move test for .CURDIR from misc.mk to varname-dot-curname.mk

(rillig)

make(1): make individual targets in test archive.mk visible

This test still fails on FreeBSD, printing the "ar cru, ranlib" lines
twice, and I don't know why.

To get more insights into the inner details, without activating the full
-dA logging, split the output into paragraphs, one per target.

This change reveals another interesting detail: Dependencies without any
commands (in this case create-archive) don't inherit the commands from
.USEBEFORE and .USE sources.

(rillig)

Tickets #1097 - #1105

(martin)

Pull up following revision(s) (requested by he in ticket #1105):

sys/arch/arm/include/arm32/vmparam.h: revision 1.56

Bump MAXDSIZ from 1536 to 1836.
1836 is 2048-128(MAXTSIZ)-64(MAXSSIZ).

Gives us a fighting chance to build rust natively.

OK'ed by skrll@

(martin)

Tickets #1611 - #1613

(martin)

Pull up following revision(s) (requested by roy in ticket #1613):

sys/netinet/in.c: revision 1.241
sys/netinet6/in6.c: revision 1.282

inet: Treat LINK_STATE_UNKNOWN as LINK_STATE_UP when changing

It's something we have always done.
it's really rare for anything to transition to UNKNOWN from either
UP or DOWN, but technically it is possible.

(martin)

Pull up following revision(s) (requested by roy in ticket #1104):

sys/netinet/in.c: revision 1.241
sys/netinet6/in6.c: revision 1.282

inet: Treat LINK_STATE_UNKNOWN as LINK_STATE_UP when changing

It's something we have always done.
it's really rare for anything to transition to UNKNOWN from either
UP or DOWN, but technically it is possible.

(martin)

make(1): fix test opt-debug-jobs on non-native platforms

(rillig)

Pull up following revision(s) (requested by knakahara in ticket #1612):

sys/netipsec/xform_esp.c: revision 1.101

Make sequence number of esp header MP-safe for IPsec Tx side. reviewed by ozaki-r@n.o

In IPsec Tx side, one Security Association can be used by multiple CPUs.
On the other hand, in IPsec Rx side, one Security Association is used
by only one CPU.

XXX pullup-{8,9}

(martin)

Pull up following revision(s) (requested by knakahara in ticket #1103):

sys/netipsec/xform_esp.c: revision 1.101

Make sequence number of esp header MP-safe for IPsec Tx side. reviewed by ozaki-r@n.o

In IPsec Tx side, one Security Association can be used by multiple CPUs.
On the other hand, in IPsec Rx side, one Security Association is used
by only one CPU.

XXX pullup-{8,9}

(martin)

Pull up following revision(s) (requested by kim in ticket #1611):

etc/ntp.conf: revision 1.21
etc/ntp.conf: revision 1.22

Use "pool" for the pool.ntp.org servers. Add some new hints.
- Use the "pool" keyword for obtaining servers from ntp.pool.org.
  - Add "tos minclock" and "tos maxclock" to limit the number of servers.
  - Add "restrict source" to apply appropriate restrictions to servers.
    (Specifically "nopeer" cannot be applied to "pool" servers.)
  - A single "pool" entry suffices -- using "2.netbsd.pool.ntp.org" so
    that we get both IPv4 and IPv6 addresses. (No addresses are returned
    for just "netbsd.pool.ntp.org.")
- Add a comment about "tinker panic 0" -- useful for VMs and laptops.
- Add a comment about "discard minimum" -- useful for some SNTP clients.
- Add an explanation for the "limited" restriction keyword.
- Unify whitespace and comment formatting.

Add iburst to peer and server.

(martin)

Pull up following revision(s) (requested by kim in ticket #1102):

etc/ntp.conf: revision 1.21
etc/ntp.conf: revision 1.22

Use "pool" for the pool.ntp.org servers. Add some new hints.
- Use the "pool" keyword for obtaining servers from ntp.pool.org.
  - Add "tos minclock" and "tos maxclock" to limit the number of servers.
  - Add "restrict source" to apply appropriate restrictions to servers.
    (Specifically "nopeer" cannot be applied to "pool" servers.)
  - A single "pool" entry suffices -- using "2.netbsd.pool.ntp.org" so
    that we get both IPv4 and IPv6 addresses. (No addresses are returned
    for just "netbsd.pool.ntp.org.")
- Add a comment about "tinker panic 0" -- useful for VMs and laptops.
- Add a comment about "discard minimum" -- useful for some SNTP clients.
- Add an explanation for the "limited" restriction keyword.
- Unify whitespace and comment formatting.

Add iburst to peer and server.

(martin)

Pull up following revision(s) (requested by nia in ticket #1101):

share/man/man8/afterboot.8: revision 1.74

ifconfig up before doing wifi stuff

(martin)

Pull up following revision(s) (requested by nia in ticket #1100):

sys/dev/pckbport/synaptics.c: revision 1.69
sys/dev/pckbport/synaptics.c: revision 1.70
share/man/man4/pms.4: revision 1.38

erect a signpost for weary travellers

synaptics: Emulate scrolling when the middle button is held with TrackPoints
idea stolen from various other operating systems.

this configurable with a sysctl in case somebody wants to hold the middle
button, e.g.  with old window managers that close menus when a button is
released.

(martin)

Pull up following revision(s) (requested by nia in ticket #1099):

external/mit/ctwm/libexec/Makefile: revision 1.1
external/mit/ctwm/bin/ctwm/Makefile: revision 1.6
external/mit/ctwm/bin/ctwm/Makefile: revision 1.7
external/mit/ctwm/bin/ctwm/Makefile: revision 1.8
external/mit/ctwm/libexec/ctwm_app_menu: revision 1.1
distrib/sets/lists/xbase/mi: revision 1.157
external/mit/ctwm/Makefile: revision 1.2
external/mit/ctwm/Makefile: revision 1.3
external/mit/ctwm/Makefile: revision 1.4
external/mit/ctwm/Makefile: revision 1.5
external/mit/ctwm/etc/system.ctwmrc: revision 1.1
external/mit/ctwm/etc/Makefile: revision 1.1
external/mit/ctwm/etc/Makefile: revision 1.2

add a script for dynamically generating a menu in the ctwm desktop
this parses the files in /usr/pkg/share/applications and provides a
submenu for launching applications.

TODO: make /usr/pkg configurable somehow?

move the new ctwmrc here so it doesn't conflict with the upstream one

revert; use system.ctwmrc from xsrc until I understand why the build fails

unbreak. bsd.files.mk needs to have an includes target, see also
share/wscons/fonts/Makefile

(martin)

openssl: Define OPENSSL_NO_EC_NISTP_64_GCC_128 for Alpha too

The ecp_nistp521.c code depends on doing unaligned 64-bit accesses,
which the NetBSD/alpha kernel doesn't like to trap and emulate.

fixes PR lib/55701

(jakllsch)

Pull up following revision(s) (requested by nia in ticket #1099-x.req):

external/mit/xdm/dist/config/Xsession.in: revision 1.2
external/mit/xdm/dist/config/Xsession.in: revision 1.3
external/mit/xdm/dist/config/Xsession.in: revision 1.4
external/mit/xdm/dist/config/Xsession.in: revision 1.5
external/mit/xinit/dist/xinitrc.cpp: revision 1.5
external/mit/xinit/dist/xinitrc.cpp: revision 1.6
external/mit/xsm/dist/system.xsm: revision 1.2
external/mit/xinit/dist/xinitrc.cpp: revision 1.7

make the default xinitrc a bit nicer

spawn only one xterm in unicode mode and use ctwm instead of twm
ctwm is a backwards compatible fork of twm that includes some extra
features like virtual desktops

spawn ctwm in the default xdm Xsession

turn off the ctwm "welcome" screen
while this is nice for Branding(tm), it adds a slight delay to startup,
which is slightly annoying, and a departure from how we've done things
previously.

update the default xsm session to ctwm/uxterm

remove redundant $

Slightly nicer default colors for various X11 apps

(martin)

Pull up following revision(s) (requested by nia in ticket #1098):

external/mit/xorg/bin/xeyes/Makefile: revision 1.3

xeyes: Enable anti-aliasing with XRENDER

If you prefer the "classic" look, "xeyes +render"

(martin)

Pull up following revision(s) (requested by nia in ticket #1097):

external/mit/xorg/bin/xterm/Makefile: revision 1.18

xterm: Enable SIXEL graphics support.

could also enable REGIS, it will require -lm and i suppose isn't
a default upstream yet(?)

test with "-ti vt340".

(martin)

Mark US/Pacific-New obsolete

(skrll)

%#jx vs %jx consistency in UVMHIST_LOG

(skrll)

tzdata 2020b: the ancient systemv and pacificnew zone files are gone,
as is the yearistype.sh script (2020b zic, if/when that gets included
in NetBSD won't support it).,

(kre)

Bump MAXDSIZ from 1536 to 1836.
1836 is 2048-128(MAXTSIZ)-64(MAXSSIZ).
Gives us a fighting chance to build rust natively.
OK'ed by skrll@

(he)

improve register descriptions

(nisimura)

PR kern/45117

Work around regression introduced in rev 1.92:

http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/kern/kern_cpu.c#rev1.92

by which ``cpuctl offline n'' became broken on architectures without
__HAVE_INTR_CONTROL (i.e., everything other than alpha and x86);
cpu_setintr() always fails on these archs, and we had neglected
return value from that function until rev 1.91.

XXX
As martin pointed out in the PR, I'm not sure whether fix in rev 1.92
itself is correct or not. Insert XXX comment referring the PR there....

(rin)

Apply hack for GDB 8.3 on arm to GDB 11.

XXX
Not tested as GDB 11 does not build for arm at the moment.

Note that dist/gdb/dwarf2foo.c was renamed to dist/gdb/dwarf2/foo.c.

(rin)

GCC 9.3 miscompiles dwarf2{expr,loc}.c with -O2 for earmv5hf{,eb}
(-O1 is fine). This is a new regression introduced in GCC9.

For everyone's safety, extend -O0 hack to dwarf2{expr,loc}.c for
all arm variants with GCC >= 8.

(rin)

Note update to tzdata 2020b

(kre)

Merge tzdata2020b

(kre)

expat-2.2.10 out.

(wiz)

- for readability when the exponent < 10 don't prefix it with 0x (from kre)
- fix usage
- merge printing code.

(christos)

Eddie Van Halen, co-founder of Van Halen dies.

(nat)

Add a new, more aggressive allocator for uvm_pglistalloc() to allocate
contiguous physical pages, and try this new allocator if the existing
one fails.  The existing contig allocator only tries to allocate pages
that are already free, which works fine shortly after boot but rarely
works after the system has been up for a while.  The new allocator uses
the pagedaemon to evict pages from memory in the hope that this will
free up a range of pages that satisfies the constraits of the request.
This should help with things like plugging in a USB device, which often
fails for some USB controllers because they can't get contigous memory.

(chs)

Only touch PMC registers if Performance Monitor Extensions are present.

(jmcneill)

- qemu_hardclock(): fix check for spurious hardclock call.
- qemu_clock_init(): Initialize qemu_nsec_per_tick *after* adjusting hz.
  Error pointed out by Jonathan Kollasch.

(thorpej)

PR port-alpha/54307

GCC 9.3 seems to be able to compile rtree.c with -O2:

- No new regressions in ATF.
- System survives over a night, at least, under heavy loads.

On the other hand, unfortunately, GCC 9.3 still miscompiles tcache.c
with -O2 or -O1. For example, even ``gcc -g hello.c'' fails with ICE
if tcache.c is compiled with -O[12] in libc.

(rin)

Comment nit

(skrll)

Comment nit

(skrll)

net80211: Initialise the interface with a decent link state.

Link state transitions to UP when a node is joined and DOWN when left.
This means that with the interface UP, the link state could be UNKNOWN
for a while, implying it can be used in BSS mode.
Which is of course false.

Add a function to set an initial link state based on the operating mode.
Also call this when the operating mode changes.

Basically in BSS and MONITOR it starts off down.
BSS will transition UP and DOWN as before, MONITOR will stay down.
IBSS, AHDEMO and HOSTAP will remain as link unknown because the state is
..... unknown.

(roy)

Replace local patches adding NetBSD/aarch64 support with the upstream one

Cherry-pick:

>From c0b313441717b65569edb01bf9984d2066d899de Mon Sep 17 00:00:00 2001
From: Nick Clifton <nickc@redhat.com>
Date: Mon, 5 Oct 2020 14:56:42 +0100
Subject: [PATCH] Add NetBSD AArch64 Little Endian and Big Endian BFD support.

bfd      * config.bfd (aarch64-*-netbsd*, aarch64_be-*-netbsd*): Add target.

(kamil)

Undo local patches

They are no longer needed (and are wrong).

(kamil)

Remove NFC local delta

(kamil)

Fix typo

Remove delta with upstream.

(kamil)

Replace local patch with fix from upstream

Cherry-pick:

>From e2a2a24a8e78427ff8667d625f5befbe88c328bb Mon Sep 17 00:00:00 2001
From: Kamil Rytarowski <n54@gmx.com>
Date: Wed, 23 Sep 2020 04:20:05 +0200
Subject: [PATCH] Preinitialize the sockaddr_un variable to zero

Don't pass random sun_len for the BSD's,
zero the whole structure as recommended for portability.

Reported by Coverity.

gdbsupport/ChangeLog:

* agent.cc (gdb_connect_sync_socket): Preinitialize addr with zeros.

(kamil)

Undo local NFC patch

(kamil)

make(1): remove incomplete fallback implementation of strftime

The function strftime is available since C89.  Any environments older
than that are not supported anymore.

(rillig)

Undo local patch having no impact

/tmp is picked from P_tmpdir earlier.

(kamil)

Undo local NFC patch

(kamil)

make(1): add test for double expansion in SySV modifier

(rillig)

make(1): migrate SysV modifier test to use the preprocessor

When both the expected and the actual expression are written in the same
line of the same file, it is easier to compare them and to document
anything interesting.  The exp file doesn't provide any space for
comments or explanations.

(rillig)

Use raw device for configuring units. This is necessary as
having a block device opened prevents autodiscovery of wedges.

(mlelstv)

Fix ioctl locking. Add dkdriver.

(mlelstv)

make(1): remove macro FILENO

This macro was obviously wrong since it would have converted file
numbers above 127 to very large numbers, close to 2^32.

The fact that it didn't blow up at all is that this macro was only ever
given the file descriptor 4 as an argument, which can well be
represented as a char, be it signed or unsigned.  And this is how the
story goes:

In Job_Init, two jobs are started.  The server job allocates file
descriptors 15 and above.  The childExitJob is created next, and the
pipe that it creates is [3, 4].  Using F_DUPFD, fd 3 is mapped to fd 5,
and fd 3 is closed.  After that, fd 4 is mapped to fd 3 (which had just
been closed), and fd 4 is closed.

After this initialization, file descriptors 0, 1, 2, 3 and 5 are taken.
This leaves a gap at file descriptor 4, and this gap is filled whenever
cmdFILE is created.

Because of this particular order of events, the macro is not necessary.
If it should ever become necessary on platforms like the old SunOS, the
parameter minfd for JobCreatePipe should be increased to 5, which would
leave the file descriptors 3 and 4 to be used by stdio streams.

On 1995-11-03, when the macro was added, SunOS must have been in its
very early development.  In Solaris 8 (released in January 2000),
FILE._file is already an unsigned char, therefore the seeming need for
this macro must have been due to an older version, probably from the 2.x
series of SunOS.

(rillig)

Check dkdriver before calling a driver function.

(mlelstv)

GC unused MAXTSIZ32

(christos)

make MAXTSIZ optional

(christos)

Make MAXTSIZ optional.

(christos)

Revert the previous again; miniroot overflows this time...

- switch to x_disklabel (drop non-native label support)
- switch to x_fsck_ffs and x_newfs (drop byte-swapped and Apple UFS support)
- switch to more

(rin)

make(1): rework memory allocation for the name of variables

There's more to know about variable names than fits in a one-liner.
While here, enforce that the name is not modified by splitting it into
the established (var + var_freeIt) pattern.

Since the name is not modified and not freed in the middle of evaluating
an expression, there is no need to make a backup copy of it.  That code
had been necessary more than 12 years ago, but not anymore since the
code got a lot cleaner since then.

(rillig)

make(1): remove duplicate code for creating variables

(rillig)

move #include "opt_compat_netbsd32.h" to where it's required

(skrll)

make(1): remove pathname limit for Dir_FindHereOrAbove

While trying to compile the code with GCC's -Wformat-truncation, the
snprintf calls felt quite complicated.  The function Dir_FindHereOrAbove
is not in a bottleneck execution path, therefore it doesn't hurt to
dynamically allocate the memory instead of using size-limited stack
memory.

(rillig)

make(1): enable format string truncation warnings for parse.c

No node is ever going to have a million or more cohorts.  And even if
there were, this name is only used for printing informational messages.

(rillig)

make(1): prepare job.c, main.c, parse.c, suff.c for WARNS=6

In job.c, GCC 5 complains about the macro FILENO that it has type
unsigned int, which is then passed as the argument of dup2, which
expects a simple int.  Maybe this workaround from 1995 is not necessary
anymore, or maybe it is but can be restricted to the few affected
platforms.

This leaves meta.c and the filemon files to be converted.  I didn't do
them since they are not well covered by the unit tests.

(rillig)

PR/55695: Andreas Gustafsson: factor(6) -h option doesn't always work
Handle -h for factors greater than the primes table.

(christos)

make(1): make dir.c, for.c and hash.c ready for WARNS=6

Some types have changed from int to unsigned int, size_t or time_t.

The variable i in hash.c has been kept as int since it counts down to
-1, which generates efficient machine code, at least on x86_64.

(rillig)

make(1): make cond.c ready for WARNS=6, part 2

(rillig)

make(1): make cond.c ready for WARNS=6

(rillig)

make(1): fix compilation with GCC 10 and -Wimplicit-fallthrough=4

(rillig)

make(1): fix double-free bug in -DCLEANUP mode (since 2020-10-02)

The bug had been introduced with dir.c 1.155 on 2020-10-02 22:20:25.  In
that commit, openDirectories was replaced with a combination of a list
with a hash table, for more efficient lookup by name.

Upon cleanup, OpenDirs_Done is called, which in turn called
Dir_ClearPath.  Dir_ClearPath takes full ownership of the given list and
empties it.  This was no problem before since afterwards the list was
empty and calling Lst_Free just frees the remaining list pointer.

With OpenDirs, this list was combined with a hash table, and the hash
table contains the list nodes, assuming that the OpenDirs functions have
full ownership of both the list and the hash table.  This assumption was
generally correct, except for the one moment during cleanup where full
ownership of the list was passed to Dir_ClearPath, while the hash table
still contained pointers to the (now freed) list nodes.  This by itself
was not a problem since the hash table would be freed afterwards.  But
as part of Dir_ClearPath, OpenDirs_Remove was called, which looked up
the freed directory by name and now found the freed list node, trying to
free it again.  Boom.

Fixed by replacing the call to Dir_ClearPath with code that only frees
the directories, without giving up control over the list.

(rillig)

make(1): revert previous commit

It had accidentally reverted all the work from the past few days.

(rillig)

make(1): fix double-free bug in -DCLEANUP mode (since 2020-10-02)

The bug had been introduced with dir.c 1.155 on 2020-10-02 22:20:25.  In
that commit, openDirectories was replaced with a combination of a list
with a hash table, for more efficient lookup by name.

Upon cleanup, OpenDirs_Done is called, which in turn called
Dir_ClearPath.  Dir_ClearPath takes full ownership of the given list and
empties it.  This was no problem before since afterwards the list was
empty and calling Lst_Free just frees the remaining list pointer.

With OpenDirs, this list was combined with a hash table, and the hash
table contains the list nodes, assuming that the OpenDirs functions have
full ownership of both the list and the hash table.  This assumption was
generally correct, except for the one moment during cleanup where full
ownership of the list was passed to Dir_ClearPath, while the hash table
still contained pointers to the (now freed) list nodes.  This by itself
was not a problem since the hash table would be freed afterwards.  But
as part of Dir_ClearPath, OpenDirs_Remove was called, which looked up
the freed directory by name and now found the freed list node, trying to
free it again.  Boom.

Fixed by replacing the call to Dir_ClearPath with code that only frees
the directories, without giving up control over the list.

(rillig)

make(1): fix test for .ifndef when compiled with -DUSE_UCHAR_BOOLEAN

In that compilation variant, TRUE is defined to 255, to see whether all
boolean expressions evaluate to either 1 or 0.  The field If.doNot in
cond.c doesn't do this since it uses the actual value of TRUE.
Therefore, change the evaluation slightly to also handle this case.

(rillig)

make(1): clean up compat.c

Only cosmetic changes, no change of functionality.

(rillig)

make(1): extract init_machine and init_machine_arch from main

(rillig)

Minor correction to prior

(roy)

make(1): extract parsing of sources from ParseDoDependency

(rillig)

make(1): extract the target parsing from ParseDoDependency

(rillig)

make(1): split ParseDoDependency into several smaller functions

(rillig)

ifconfig: Warn once more if media supported but no types

This reverts media.c -r1.7

(roy)

ppp: Remove media

There is none after all.
Applications should be using ifi_link_state and not checking media.

(roy)

Note import of dhcpcd-9.3.0

(roy)

Sync with dhcpcd-9.3.0

(roy)

make(1): split ParseDoSrc into smaller functions

(rillig)

make(1): replace GNode macros with inline functions

Curiously, this affects the size of the generated code: it gets smaller
than before.

(rillig)

make(1): allow easy creation of preprocessed source and assembler

To be used during development, to see how changes in the source code
affect the generated machine code.

(rillig)

NGPC

(nia)

bump date

(nia)

revert previous octal conversion code (broken and not worth it).

(christos)

revert previous and don't parse octal

(christos)

Update gdb/dwarf2expr.c hack for earmv7hf{,eb}; the situation does not
change even for GCC 9.3 with ``memcmp'' fix.

(rin)

Update aes_ccm_tag() hack for m68k.

Even with ``memcmp fix'', GCC 9.4 miscompiles this function for -O[12].
But the situation was slightly changed from that with GCC 8.3:

* -O0 and -O1 work but -O2 fails for 68060 and 68040 (real hardware)
* -O0 and -O2 work but -O1 fails for 68020 and 68010 (TME)

(rin)